backup/warcprox
1
0
Fork 0
mirror of https://github.com/internetarchive/warcprox.git synced 2025-01-18 13:22:09 +01:00
Code Issues Projects Releases Activity

add payload digest header to revisit records

Browse Source
This commit is contained in:
Noah Levitt 2015-03-26 15:17:46 -07:00
parent 0eb2917e50
commit 965853f4ab
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
warcprox/warcwriter.py
View File

@ -75,6 +75,7 @@ class WarcWriter:
refers_to=dedup_info['i'],
refers_to_target_uri=dedup_info['u'],
refers_to_date=dedup_info['d'],
payload_digest=self.digest_str(recorded_url.response_recorder.payload_digest),
profile=warctools.WarcRecord.PROFILE_IDENTICAL_PAYLOAD_DIGEST,
content_type=hanzo.httptools.ResponseMessage.CONTENT_TYPE,
remote_ip=recorded_url.remote_ip)
@ -104,7 +105,7 @@ class WarcWriter:
def build_warc_record(self, url, warc_date=None, recorder=None, data=None,
concurrent_to=None, warc_type=None, content_type=None, remote_ip=None,
profile=None, refers_to=None, refers_to_target_uri=None,
refers_to_date=None):
refers_to_date=None, payload_digest=None):
if warc_date is None:
warc_date = warctools.warc.warc_datetime_str(datetime.utcnow())
@ -131,6 +132,8 @@ class WarcWriter:
headers.append((warctools.WarcRecord.CONCURRENT_TO, concurrent_to))
if content_type is not None:
headers.append((warctools.WarcRecord.CONTENT_TYPE, content_type))
if payload_digest is not None:
headers.append((warctools.WarcRecord.PAYLOAD_DIGEST, payload_digest))
if recorder is not None:
headers.append((warctools.WarcRecord.CONTENT_LENGTH, str(len(recorder)).encode('latin1')))